Handling of ocean runoff points for CESM coupled configurations

[ekluzek]

We realized that for coupling to CESM we need mizuRoute to handle ocean points. Specifically it needs to handle points that CTSM thinks are land -- but mizuRoute thinks are ocean. This of course is around coastal areas, and there is significant area over Antarctica where this is the case. MOSART handles this because it has a global grid that includes non-land points.

To do this for mizuRoute we figure the ocean gridcells can just be the MOSART half degree gridcells and they can be "tacked onto the end" of the regular mizuRoute grid that includes the land gridcells for mizuRoute. By doing it this way mizuRoute code shouldn't have to be touched and only the top level cap would have arrays of a larger size.

Since the data handling for these ocean points is minimal it's probably OK to start with having just the main task (MPI rank 0) handle the ocean cells. This will make the handling pretty straight forward. Later this could be extended to handle some ocean cells on other tasks.

Design Goals:

• Handle this in the NUOPC cap (under the route/build/cpl directory)
• Don't change any mizuRoute standalone code
• Leave the network input dataset alone
• Start with a simple half degree grid over ocean points
• Start with doing the ocean handling just in the main task

Some particulars:

• We'd want this to handle both the current datasets that work with standalone CTSM "I compset" and these new ones needed for fully coupled
• The mesh files would change
• The mapping files would change
• Area for the ocean cells will need to be on the MESH file
• Mask will need to be added to the MESH file